RRT

Table of contents

1. Core idea
2. Why use RRT
3. Common variants

---

Rapidly-exploring Random Trees (RRT) are sampling-based motion planning methods that incrementally grow a tree in configuration space.

Core idea

RRT expands from a start state by repeatedly:

1. Sampling a random state.
2. Finding the nearest node already in the tree.
3. Extending toward the sample with a bounded step.
4. Accepting the new node if it is collision-free.

Why use RRT

• Works well in high-dimensional spaces.
• Handles complex obstacle geometries.
• Produces feasible paths quickly.

Common variants

• RRT-Connect
• RRT*
• Informed RRT*